dataSDA 0.2.7

Correct or fix the references/citations/source of the datasets.

dataSDA v0.2.6

dataSDA v0.2.6 — minor update of the CRAN package (previous CRAN version 0.2.5).

R CMD check --as-cran (R 4.6.0, Windows): 0 errors | 0 warnings | 0 notes.

New features

• check_zero_width_intervals() — diagnostic that flags zero-width
  intervals (min == max) in interval-valued data. Accepts both MM format
  (paired _min/_max columns) and RSDA format (symbolic_tbl with
  symbolic_interval columns). Returns (invisibly) a logical scalar carrying
  a "flagged" matrix and a "variables" vector, and warns by default.
  A tol argument allows near-zero widths to be flagged. Useful for screening
  data before tools that divide by interval width (e.g.
  ggInterval::ggInterval_indexImage()).

• aggregate_to_symbolic(type = "int") gains a zero_width argument
  controlling how zero-width intervals in the aggregated output are handled:

  • "keep" (default) — leave the output untouched; zero-width intervals
    are returned as-is, with no removal, no adjustment, and no warning.
  • "remove" — drop every concept containing a zero-width interval.
  • "regenerate" — re-run the aggregation until none remain (stochastic
    group_by only).
  • "adjust" — add a small epsilon to each zero-width upper endpoint.

  A companion epsilon argument sets the adjustment amount. For "remove"
  and "adjust", a single warning names the affected variables and the action
  taken.

Other changes

• environment.mix: the four modal columns are restored to proper
  symbolic_modal vctrs so the object matches ggInterval::Environment in
  class and format; documentation updated accordingly. Regeneration script
  added under data-raw/.
• Vignette polished (subtitle, ggInterval plotting calls, v0.2.6 citation).
  The eval = NOT_CRAN gate keeps the CRAN vignette build fast (code-only);
  full evaluated renders are attached below as assets.
• Stopped tracking generated inst/doc/ artifacts so built tarballs no longer
  ship orphaned vignette figures.

See NEWS.md for the complete changelog.

Assets

• dataSDA_0.2.6.tar.gz — the CRAN source tarball (passes R CMD check --as-cran).
• dataSDA_intro.html / dataSDA_intro.pdf — fully evaluated vignette
  (rendered offline with NOT_CRAN=true).

dataSDA v0.2.5

What's New in v0.2.5

New Datasets

• 11 interval time series (ITS) datasets: crude_oil_wti.its, djia.its, euro_usd.its, ibovespa.its, irish_wind.its, merval.its, petrobras.its, shanghai_stock.its, sp500.its

New Functions

• aggregate_to_symbolic() — convert traditional data to symbolic data format
• read_symbolic_csv() / write_symbolic_csv() — read and write symbolic data as CSV
• search_data() — search available datasets by keyword or type
• to_all_interval_formats() — convert intervals to all supported formats
• ARRAY format converters: ARRAY_to_MM, ARRAY_to_RSDA, ARRAY_to_iGAP, MM_to_ARRAY, RSDA_to_ARRAY, SODAS_to_ARRAY, iGAP_to_ARRAY

Dataset Renaming

• Datasets renamed with proper type suffixes for consistency (e.g., crime → crime.modal, mushroom → mushroom.int.mm, mushroom_fuzzy → mushroom_fuzzy.mix)

Bug Fixes & Improvements

• Fixed int_dist p parameter bug
• Updated documentation for all datasets and functions
• Expanded vignette with new examples and content

dataSDA v0.2.4

What's Changed

• Fix documentation for 19 datasets with incorrect column metadata: Corrected @format sections for datasets stored in Min-Max format (bats.int, oils.int, teams.int, tennis.int, trivial_intervals.int, credit_card.int, temperature_city.int, bird_species.mix, employment.int, town_services.mix) and fixed wrong column names/counts in cars.int, china_temp.int, environment.mix, health_insurance.mix, horses.int, nycflights.int, synthetic_clusters.int, uscrime.int, china_temp_monthly.int
• All 105 datasets now have complete, accurate @format documentation with correct dimensions, column names, and type descriptions

Full Changelog: v0.2.3...v0.2.4

dataSDA v0.2.3

What's Changed

• Complete documentation for all 105 datasets: Added proper @format sections with variable names, types, dimensions, and descriptions to 29 previously under-documented datasets
• Fixed 3 datasets with incorrect metadata: bird_species_extended.mix, world_cup.int, acid_rain.int now have accurate column counts and variable descriptions
• Added cross-references between related dataset pairs (wide/modal formats, iGAP/MM formats)

Full Changelog: v0.2.2...v0.2.3

dataSDA v0.2.2

New datasets (17 added, 105 total)

17 new datasets from R packages (intkrige, RSDA, MAINT.Data, GPCSIV, HistDAWass, symbolicDA), the Billard & Diday (2007) textbook, and the QualAr Portuguese air quality network.

Interval-valued (12)

• utsnow.int — 415 Utah weather stations with snow load prediction intervals (from intkrige)
• lynne1.int — 10 observations with pulse/blood pressure intervals (from RSDA)
• loans_by_risk_quantile.int — 35 Lending Club loan groups with quantile-based intervals (from MAINT.Data)
• judge1.int, judge2.int, judge3.int — 6 regions × 4 interval variables (from GPCSIV)
• video1.int, video2.int, video3.int — 10 user groups × 5 engagement intervals (from GPCSIV)
• lisbon_air_quality.int — 1096 daily observations of 8 pollutant intervals (QualAr)

Histogram-valued (4)

• blood.hist — 14 gender-age groups with 3 blood test histograms (from HistDAWass)
• china_climate_month.hist — 60 stations × 168 monthly climate histograms (from HistDAWass)
• china_climate_season.hist — 60 stations × 56 seasonal climate histograms (from HistDAWass)
• exchange_rate_returns.hist — 108 monthly return histograms (from HistDAWass)

Mixed symbolic (1)

• polish_cars.mix — 30 car models with 9 interval + 3 multinomial variables (from symbolicDA)

From Billard & Diday (2007) (2)

• hierarchy.hist — 10 observations with hierarchical categories and conditional histograms (Table 6.20)
• bird_color_taxonomy.hist — 20 birds with density/size histograms and fuzzy shade taxonomy (Tables 6.9/6.14)

Quality

• All 517 tests pass
• R CMD check: 0 errors, 0 warnings, 0 notes

v0.2.1: 19 new datasets (88 total)

New datasets (19 added, 88 total)

19 new datasets from Billard & Diday (2020) Clustering Methodology for Symbolic Data, the HistDAWass R package, and other R packages.

Interval-valued (5)

• genome_abundances.int — 14 genome classes × 10 dinucleotide abundance intervals
• china_temp_monthly.int — 15 Chinese weather stations × 12 monthly temperature intervals
• ecoli_routes.int — 9 E. coli transport routes × 5 interval variables
• loans_by_risk.int — 35 loan groups by risk level × 4 financial intervals (from MAINT.Data)
• polish_voivodships.int — 18 Polish voivodships × 9 socio-economic intervals (from clusterSim)

Histogram-valued (9)

• iris_species.hist — 3 iris species × 4 morphological histograms
• flights_detail.hist — 16 airlines × 5 flight performance histograms
• cover_types.hist — 7 forest cover types × 4 topographic histograms
• glucose.hist — 4 regions × 1 blood glucose histogram
• state_income.hist — 6 US states × 4 income histograms
• simulated.hist — 5 observations × 2 simulated histograms
• age_pyramids.hist — 229 countries × 3 age pyramid histograms (from HistDAWass)
• ozone.hist — 84 daily observations × 4 weather histograms (from HistDAWass)
• french_agriculture.hist — 22 French regions × 4 agricultural histograms (from HistDAWass)

Distribution-valued (1)

• household_characteristics.distr — 12 counties × 3 categorical distributions

Mixed symbolic (4)

• county_income_gender.hist — 12 counties with gendered income histograms + sample sizes
• joggers.mix — 10 jogger groups with interval + histogram variables
• census.mix — 10 census regions with 6 mixed-type variables
• mtcars.mix — 5 car groups with 7 interval + 4 modal variables (from ggESDA)

dataSDA 0.2.0

New datasets (13 added, 69 total)

From R packages

• cardiological.int — 44 patients, 5 interval physiological measurements (RSDA)
• prostate.int — 97 prostate cancer patients, 9 clinical interval variables (RSDA)
• uscrime.int — 46 US states, 102 interval crime statistics (RSDA)
• hardwood.hist — 5 hardwood species, 4 histogram-valued climate variables (RSDA)
• synthetic_clusters.int — 125 observations in 5 clusters, 6 interval variables (symbolicDA)
• environment.mix — 14 EPA state groups, mixed interval/modal environmental data (ggESDA)

From Billard & Diday (2006) textbook

• weight_age.hist — 7 age groups, weight histograms (Table 3.10)
• hospital.hist — 15 hospitals, cost histograms (Table 3.12)
• cholesterol.hist — 14 gender-age groups, cholesterol histograms (Table 4.5)
• hemoglobin.hist — 14 gender-age groups, hemoglobin histograms (Table 4.6)
• hematocrit.hist — 14 gender-age groups, hematocrit histograms (Table 4.14)
• hematocrit_hemoglobin.hist — 10 bivariate 2-bin histograms (Table 6.8)
• energy_usage.distr — 10 towns, categorical fuel/heating distributions (Table 3.7)

v0.1.9

New datasets (7 added, 55 total)

Seven new interval-valued benchmark datasets from recent SDA papers (2020-2025):

• freshwater_fish.int — 12 freshwater fish species, 13 heavy metal bioaccumulation variables, 4 feeding classes
• fungi.int — 55 fungi specimens, 5 morphological variables, 3 genera (Amanita/Agaricus/Boletus)
• iris.int — 30 interval observations of Fisher's iris data, 4 sepal/petal variables, 3 species
• water_flow.int — 316 water flow sensor readings, 47 interval features, 2 classes
• wine.int — 33 wine samples, 9 chemical/physical property variables, 2 classes
• car_models.int — 33 Italian car models, 8 specification variables, 4 categories
• hdi_gender.int — 183 countries, 2 World Bank gender indicator intervals, ordinal HDI classification

Sources

• Andrade, N. A. et al. (2025). Kernel clustering with automatic variable weighting for interval data. Neurocomputing, 617, 128954.
• Alcacer, A. et al. (2023). Ordinal classification for interval-valued data. Expert Systems with Applications, 238, 121825.

Verification

• R CMD check: 0 errors, 0 warnings, 0 notes
• All 517 existing tests pass

dataSDA v0.1.8

dataSDA 0.1.8

Vignette

• Comprehensive rewrite covering all 51 exported functions (was 18), organized into 14 sections: format detection/conversion, core stats, geometry, position, robust, shape, similarity, uncertainty, distance, histogram stats, and utilities.
• Fixed garbled characters, stale dataset references, and broken histogram example code.
• Updated method descriptions in vignette.
• Added pre-built vignette to inst/doc/ to fix CRAN check WARNING.

CRAN

• 0 errors, 0 warnings, 0 notes on R CMD check --as-cran.